Fig. 5
The influences of KL-VShet genotype (KL-VShet− vs. KL-VShet+) on longitudinal tau propagations. Negative interaction of KL-VShet− × Aβ in longitudinal tau propagation to A fusiform and B inferior temporal region. Negative interaction of KL-VShet− × entorhinal FTP SUVR in longitudinal tau propagation to C fusiform and D inferior temporal region. Possible pathways among KL-VShet, baseline Aβ Centolids, entorhinal tau concentration, and longitudinal tau accumulations in E fusiform and F inferior temporal region. ① β1 × β2 × β3, ② β1 × β5, and ③ β4 × β3 represented three indirect pathways, and β6 + ① + ② + ③ represented the total effect. All variables have been converted to standard z-scores. Total, direct, and indirect associations were calculated based upon a 5000-iteration bootstrapping procedure. Linear regression lines and datapoints of KL-VShet− and KL-VShet+ participants were colored in gray and orange, and the control, A+/T−, and A+/T+ participants were represented by circles, triangles, and squares, respectively. Linear model fits were indicated with 95% confidence intervals. Significant p values at p < 0.05 and associated βstd were marked in bold. Solid and dashed lines showed significant and insignificant associations, and the significant indirect pathways were characterized by bold orange lines